modeling scintillation effects on l1/l5 sbas availability ...€¦ · for l1(c/a)/l5 dual-frequency...

© 2008 The MITRE Corporation. All rights reserved.

Modeling Scintillation Effects on L1/L5 SBAS Availability in North and

South America

The Atmosphere and its Effect on GNSS Systems

14 to 16 April 2008Santiago, Chile

Dr. M. Bakry El-Arini


233 of 301

Overview

• Background

• Methodology for Modeling Ionospheric Scintillation Effects in MITRE’s SBAS Availability Models

• Example Results of Modeling Ionospheric Scintillation Effects on Availability for L1/L5 SBAS in North America for nonprecision approach

• Example Results of Modeling Ionospheric Scintillation Effects on Availability for L1/L5 SBAS in North America for LPV approach


234 of 301

Summary of the Process

Scintillation Data generated by a model such as WBMOD (USAF/NWRA) or GISM (ITU) or Measurements (for each LOS for every epoch for each user location )

L1 and L5 Receiver Models:Check if the signal loses lock for each LOS based on Elevation Angle, C/N0, and Scintillation parameters (S4, σφ, p, T)

Availability Models:NPA Model (GATOM)PA Model (SWAT)

Acronyms:

GATOM: GPS Air Traffic Operations ModelGISM: Global Ionospheric Scintillation Model ITU: International Telecommunication UnionLOS: Line Of SightNWRA: North West Research Associates, Seattle, WA, USASWAT: SBAS Worldwide Availability ToolUSAF: United States Air ForceWBMOD: Wide Band Model


235 of 301

Methodology for Modeling Ionospheric Scintillation Effects in MITRE’s SBAS

Availability Models

Originally Prepared by Dr. M. Bakry El-Arini, Dr. Chris Hegarty

Dr. Michael Tran, Ing. J. P. Fernow


236 of 301

Assumptions

• SBAS user equipment will NOT likely be required to support a mixed mode– E.g., dual-frequency measurements from satellites and single-

frequency measurements from other satellites• Some user equipment might support this mode, but likely not a

minimum requirement

• MITRE availability models model the user equipment as one of three cases– All satellites, including GEOs, are dual-frequency– All satellites are L1 only– All satellites are L5 only

• Notes– Future user equipment requirements might include additional

cases– Ground reference stations are dual-frequency for all cases


237 of 301

A Simple Logic for Availability Calculation for L1(C/A)/L5 Dual-Frequency SBAS 1 of 2

• Calculate availability for L1/L5 dual-frequency user and WRSs without scintillation in North America – Normal mode is both L1(C/A) and L5 are locked for all line

of sights (LOSs) to satellites visible from the service volume

– This is used as a baseline

• Calculate availability with different levels of scintillation (see next few charts for details)– WAAS Reference Stations (WRSs)

• Dual-frequency (DF) mode for each LOS to GPS satellites• If lock is lost on either L1 or L5 or both, then don’t use this

LOS in computing HPL or VPL if the satellite is a GPS satellite

• WRSs use single-frequency (SF) GEOs


238 of 301

A Simple Logic for Availability Calculation for L1(C/A)/L5 Dual-Frequency SBAS 2 of 2

• SBAS user equipment– Pure L1/L5 dual-frequency (DF) mode

• If either L1 or L5 or both is/are lost for a particular satellite LOS, then don’t use this satellite when computing HPL or VPL

– L1 or L5 single frequency (SF) modes


239 of 301

L1/L5 GPS Dual-Frequency User Error Model (Without Scintillation)

• Chart shows different components of user equipment dual-frequency range error


240 of 301

SBAS L1/L5 Dual-Frequency User Error Model

• These charts show different components of SBAS user equipment dual-frequency range error at two fixed User Differential Range Error (UDRE) as examples

• Error for Fast and long-term correction (σflt) = UDRE*δUDRE/3.29

• (Assume UDRE = 4.5m & 6m, δUDRE = 1)


241 of 301

Single-Frequency Mode User Equipment Receiver Error Models

• L1 single-frequency mode user equipment receiver error model– RTCA DO-229D for airborne receiver

• L5 single-frequency mode user equipment receiver error model– Similar algorithms to L1 as in RTCA DO-229D for airborne

receiver– Some differences due to larger ionospheric delay and error;

see next page


242 of 301

A Simple Logic for Loss of Lock for L1 C/A GPS and GEO Satellites

• For each WRS and user location– For every snapshot, every line of sight (LOS) for every visible GPS

and GEO satellites, calculate IPP location and elevation angle– Calculate C/N0C/A(GPS) or C/N0C/A(GEO) – Calculate scintillation parameters at L1 (1575.42 MHz) using a

scintillation model (WBMOD) (Measurements can be used if available)

• Amplitude scintillation: S4(L1)• Phase scintillation: σφ(L1), T(L1) (strength of power spectral density at 1 Hz), p(L1)

(slope of power spectral density)

– Calculate Carrier Tracking error (σφε(L1))

– If σφε> threshold, declare loss of lock for this LOS. This LOS should

not be used in any availability calculations– If σφε

≤ threshold, then this LOS did not lose lock, calculate code tracking error due to scintillation and RSS it with other errors. We will assume this code tracking error = 0 in this initial phase of implementation)


243 of 301

A Simple Logic for Loss of Lock for L5 GPS and GEO Satellites

• Similar to L1 with the following additional assumptions– GPS: assume 50% power in the Q-channel and 50% in the I-

channel– GEO: assume 100% power in the I-channel


244 of 301

Receiver Lock Loops Thresholds (Degrees)

L1 (C/A) L5

Ground Rx Airborne Rx Ground Rx Airborne Rx

GPS 10.5 10.0 19.7 18.5

GEO 10.5 10.0 10.5 10.0


245 of 301

Examples of Modeling Ionospheric Scintillation Effects on Availability for L1/L5

SBAS in North America: NPA Case

Originally Prepared by Ing. Tom Hsiao


246 of 301

Overall Assumptions

• 24 GPS Martinez constellation

• 2 GEOs at 133oW and 107oW

• 38 L1/L5 WAAS Reference Stations in North America, Hawaii, and Puerto Rico

• IFOR threshold outage probability statistics for GPS– No failures on operating GEO satellites

• HAL = 556m (NPA), URA = 4.5m

• 24-hour average with 1-minute sampling interval

• User grid 2 x 2 degrees


247 of 301

Scintillation Assumptions

• Date: September 15 (Day of Year = 258)– In equatorial region, worst scintillation occurs during

Equinox months

• Sunspot number (SSN) = 150 (≈ peak of solar cycle)

• KP geomagnetic index = 1 (quiet ionosphere, no storm)

• WBMOD (V15.03, dated July 8, 2005) is used to generate scintillation parameters for each line of sight– The 95th percentile values of S4 (a measure of amplitude

scintillation) and σφ (a measure of phase scintillation) are generated

• A receiver model determines whether the generated S4 and σφ values cause a loss of lock on the satellite signal– The satellite is not used in the position solution is due to

the loss of lock


248 of 301248

L1/L5 Dual-Frequency

No Scintillation 95th Percentile Scintillation without Reacquisition

95th Percentile Scintillation with Reacquisition

These results may indicated lower availability than will be experienced in practice, since WBMODdoes not model the patchiness of ionospheric scintillation.


249 of 301249

L1 Single-Frequency





250 of 301250

L5 Single-Frequency





251 of 301

Observations

• As expected, in the absence of scintillation, the NPA availability provided by pure dual-frequency mode is higher than the availability provided by either pure single-frequency L1 mode or L5 mode

• With scintillation, the availability of NPA service provided by pure dual-frequency mode is worse than the availability of L1 or L5 single-frequency mode in the affected areas (anomaly regions)– At a user location, for any line of sight (LOS) and at any

snapshot, if the dual-frequency receiver loses lock on one or both frequencies, this LOS is not used in the position calculation.


252 of 301

Example Results of Modeling Ionospheric Scintillation Effects on Availability for L1/L5

SBAS in North America: LPV Case

Originally Prepared by Ing. Deihim Hashemi

Ing. Michael P. McLaughlinIng. Pauline Yen


253 of 301

Overall Assumptions1 of 2

• 24 GPS Martinez constellation – broadcasting L1/L5

• 2 GEOs at 133º W and 107º W– Dual-Frequency L1/L5– 4.5 meters UDRE (constant)

• 38 WAAS reference stations (WRS) – L1/L5 compatible

• Reliability Models– GPS: IFOR Threshold probability statistics– GEO: No Failures


254 of 301

Overall Assumptions2 of 2

• LPV Alert Limits– Vertical 50 meters/Horizontal 40 meters

• 24-hour average with 5-minute sampling interval

• SWAT does not model ionospheric storms– Contribution to LPV unavailability is estimated to be 0.00075

at middle latitudes in the US

• LPV results may show lower availability that will be experienced in actual practice– WBMOD does not account for patchiness of ionospheric

scintillation– SWAT does not model reacquisition


255 of 301

Ionospheric Constraint Comparison

• Case 1: No Constraint– All IPPs, whether generated from WRS-Satellite or User-

Satellite pairings, will be included in all calculations

• Case 2: Existing Slant-Line Constraint– See slide 26 for description

• Case 3: Scintillation Model Constraint– See slide 17 for scintillation assumptions– Slant-Line constraints are removed from this case

• Analysis will be conducted using varying user equipment frequency capability– L1 Single-Frequency– L5 Single-Frequency– L1/L5 Dual-Frequency


256 of 301

Slant-Line Model Constraint

Line B

S4 is the normalized standard deviation of the variation of the intensity of the received signal. S4 = 0 means no scintillation, S4 = 1.0, means severe scintillation. Severe scintillation can happen after local sunset during the peak of solar cycle.

Line A

A: Lato = (-0.175 * Lono) + 4o

• Apply to WRS-generated IPPs fromreference stations in Mexico

• Discard these IPPs from planar fit ofall IGPs in the mask (i.e. nocontribution to GIVE algorithm)

• UDRE algorithm unaffected

B: Lato = (-0.175 * Lono) – 2o

• Apply to IGPs in the mask• Raise the GIVEFloor value of those

IGPs from 3 meters to 15 meters


257 of 301

L1/L5 Dual-Frequency SBAS User Equipment Mode Full Comparison

No Ionospheric Constraints

Slant-Line Model without Scintillation Scintillation without Slant Line (95th Percentile)


258 of 301

L1 Single-Frequency SBAS User Equipment Mode Full Comparison

No Ionospheric Constraints

Slant-Line Model without Scintillation Scintillation without Slant Line (95th Percentile)


259 of 301

L5 Single-Frequency SBAS User Equipment Mode Full Comparison

No Ionospheric ConstraintsSlant-Line Model without Scintillation Scintillation without Slant Line (95th Percentile)


260 of 301

Additional Analysis for theNorthern Equatorial Region


261 of 301

Adjustments to Analysis Assumptions

• Change to Release 8/9 system architecture– Universal IGP mask (corrections available for all IGPs)

• Ionospheric constraint scenarios– No constraint– Scintillation Model constraint

• Analysis region– Central America– Caribbean– Northern South America (down to equator)


262 of 301

L1/L5 Dual-Frequency SBAS User Equipment ModeNo Ionospheric Constraints

Without Scintillation With 95th Percentile Scintillation


263 of 301

L1 Single-Frequency SBAS User Equipment ModeNo Ionospheric Constraints Modeled



264 of 301

L5 Single-Frequency SBAS User Equipment ModeNo Ionospheric Constraints Modeled



265 of 301

Acknowledgments

• Ing. Deane Bunce of the U.S. FAA for sponsoring this work

• Ing. Keith Groves of the U.S. Air Force Research Laboratory (AFRL) for providing the latest version of the scintillation model (WBMOD, Version 15)

modeling scintillation effects on l1/l5 sbas availability ...€¦ · for l1(c/a)/l5 dual-frequency...

Documents